The effect of mindfulness-based cognitive therapy on rumination and a task-based measure of intrusive thoughts in patients with bipolar disorder

Bipolar disorder (BD) is characterized by recurrent depressive and (hypo)manic episodes and substantial disability, with residual symptoms persisting in many patients despite existing treatments. A key maladaptive cognitive process in BD is depressive rumination, which is linked to depressive symptoms and relapse. Mindfulness-Based Cognitive Therapy (MBCT), effective in major depressive disorder, is hypothesized to reduce depressive rumination by increasing awareness and disengagement from automatic maladaptive cognitive patterns. The study’s primary objective was to assess whether MBCT, added to treatment-as-usual (TAU), reduces trait depressive rumination and state negative intrusive thoughts (via the Breathing Focus Task) in BD patients, and secondarily to explore effects on positive rumination and positive intrusive thoughts.

Prior research indicates MBCT reduces depressive rumination in MDD, potentially mediating symptom improvement. Evidence in BD is limited but suggests possible benefits: an RCT observed a trend toward reduced rumination post-MBCT, and an open-label trial reported significant reductions. Systematic reviews highlight MBCT as promising for BD and identify rumination as a therapeutic target. Experimental measures, such as the BFT, have been used to capture state negative intrusive thoughts, though their relation to trait rumination is not consistently observed.

Design: Substudy within a larger randomized, multicenter, evaluator-blinded clinical trial evaluating MBCT plus treatment-as-usual (TAU) versus TAU in BD, with assessments up to 15 months; current analyses used baseline and post-treatment data. Participants: Adults (≥18) with SCID-confirmed bipolar I or II; at least two lifetime depressive episodes and one affective episode in the year prior to baseline; not currently (hypo)manic (YMRS <12). Exclusions included recent manic episode (<3 months), schizophrenia/schizoaffective disorder, current substance abuse, organic brain syndrome, antisocial or borderline personality disorder, suicide or aggression risk, and medical conditions impeding participation. Recruitment: Seven specialized outpatient clinics; verbal consent followed by eligibility screening and detailed information; interested patients underwent a research interview (T0) and completed self-report questionnaires (T0a) at home. Measures: Observer-rated depressive symptoms (IDS-C), (hypo)manic symptoms (YMRS); trait depressive rumination via RRS-EXT brooding subscale; positive rumination via RPA-NL self-focused and emotion-focused subscales; experimental state intrusive thoughts via Breathing Focus Task (BFT). BFT procedure: Practice focusing on breathing (20 s) followed by a 45 s phase with three randomized tones (10–20 s intervals) to report focus or intrusive thoughts categorized as negative, positive, or neutral. Actual task: 5 min breathing with 12 randomized tones (20–30 s intervals). Post-task VAS ratings for time distracted by ruminating thoughts, positive thoughts, time focused on breathing, and difficulty focusing. Intervention: MBCT+TAU versus TAU alone. MBCT adapted from the relapse prevention manual for unipolar depression to BD needs; 8 weekly sessions (2.5 h), one 6 h silent day, daily home practice (~45 min); delivered by a BD-informed therapist and an advanced MBCT teacher per Good Practice guidelines. Timing: Baseline cognitive tasks (T0b) typically ~13 days after interview (T0) and ~9 days after self-report (T0a); post-treatment tasks (T1b) ~19.5 days after interview (T1) and ~17.5 days after self-report (T1a); T1 occurred ~18.7 weeks after T0. Sample for experimental tasks: Of 74 invited, 52 performed baseline tasks; BFT data available for 49 at baseline (MBCT+TAU N=25; TAU N=24). Pre–post BFT completed by 34, with 32 attending ≥4 MBCT sessions; final per-protocol sample for pre–post analyses: 30 participants (MBCT+TAU N=15; TAU N=15). Statistical analysis: Per-protocol pre–post analyses including MBCT participants with ≥4 sessions. Group comparisons at baseline using t-tests, Mann–Whitney, Fisher’s Exact, and χ² tests. Spearman correlations examined associations between baseline negative intrusive thoughts and depressive symptoms (IDS-C), trait rumination (RRS brooding), and self-reported time distracted by ruminating thoughts; ANOVA tested whether timing differences between T0/T0a/T0b influenced associations. Effects of MBCT on depressive and manic symptoms assessed via repeated measures ANOVA (time × group). Effects on rumination and intrusive thoughts assessed via bootstrap linear regression on change scores (post–baseline) with group as predictor; 5000 bootstrap samples with BCa 95% CIs; Cohen’s f² calculated from R². Exploratory analyses evaluated associations of positive intrusive thoughts with manic symptoms and positive rumination, and MBCT effects on positive rumination via bootstrap regressions.

Baseline associations: Negative intrusive thoughts correlated with self-reported time distracted by ruminating thoughts (rs(44)=0.53, p=0.001). Trait rumination correlated with depressive symptoms (rs(43)=0.57, p<0.0001). No significant correlations between negative intrusive thoughts and trait rumination (rs(43)=0.25, p=0.10) or depressive symptoms (rs(47)=0.17, p=0.25). MBCT effects: No significant benefit on depressive symptoms (Time × Group: F(1,28)=0.72, p=0.40, η²=0.025). Trait depressive rumination decreased in MBCT+TAU versus TAU (ΔRRS brooding ≈ −2.05 points; R²=0.16, F(1,27)=5.15, p=0.031; medium effect size f²=0.19). Negative intrusive thoughts decreased (R²=0.15, F(1,28)=4.88, p=0.036; f²=0.17). Neutral intrusive thoughts increased (R²=0.15, F(1,28)=5.09, p=0.032; f²=0.18). No effect on positive intrusive thoughts (R²=0.004, F(1,28)=0.11, p=0.74; f²=0.004) or total intrusive thoughts (R²=0.016, F(1,28)=0.45, p=0.51; f²=0.016). Exploratory outcomes: No beneficial effect on manic symptoms (Time × Group: F(1,28)=0.023, p=0.88, η²=0.001). No effect on positive rumination (RPA_ER: R²=0.002, F(1,27)=0.45, p=0.83; RPA_SR: R²=0.093, F(1,27)=2.78, p=0.11). Positive intrusive thoughts correlated with self-reported time distracted by positive thoughts (rs(44)=0.46, p=0.001), but not with manic symptoms or positive rumination.

MBCT reduced both trait depressive rumination and state negative intrusive thoughts in BD patients, supporting the hypothesis that MBCT modifies dysfunctional cognitive processes central to depression risk. The inclusion of an experimental state measure (BFT) complements self-report trait measures by capturing real-time negative intrusions, which may be less prone to recall bias. Notably, reductions occurred in a relatively euthymic sample without concurrent changes in depressive symptoms, suggesting potential preventive benefits: lower rumination may reduce risk of depressive relapse or recurrence in BD. The lack of correlation between the BFT’s negative intrusive thoughts and trait rumination suggests these measures tap distinct constructs—state-dependent intrusions versus dispositional rumination—each providing unique information. MBCT did not alter positive rumination or positive intrusive thoughts, potentially reflecting floor effects in manic symptoms or limitations of the BFT in capturing (hypo)manic cognitions. Overall, findings align with prior evidence from MDD and emerging BD studies indicating MBCT’s promise for targeting rumination in bipolar disorder.

MBCT, as an adjunct to usual care, may reduce depressive rumination and negative intrusive thoughts in BD, targeting a known risk factor for depressive relapse. State and trait measures of negative thinking appear to offer complementary insights. Future well-powered, longitudinal studies with multiple time points and refined experimental tasks should test whether reductions in rumination mediate improvements in depressive symptoms and decrease relapse risk, and clarify the predictive value of state measures.

Relatively small sample size limits power and generalizability; assessments (clinical interview, self-report questionnaires, BFT) were not conducted on the same day, potentially weakening cross-sectional associations; absence of an active control group limits attribution of effects to specific components of MBCT; BFT outcomes exhibited zero-inflation and low ranges, possibly constraining sensitivity and correlations with trait measures; pre–post design precludes causal inferences about mediation or relapse prevention; the BFT may benefit from reintroducing mood/worry induction or extending task duration; potential reinterpretation of thought valence after MBCT may influence BFT categorization.